jamovi

Hi,

I have a rather large dataset consisting of ~54 000 observations.
I try to fit a linear model with 1 dependent and 6 predictors (all scale).

It works well with the linear model module and also with Base R.
But when I try the Regression built in jamovi I get the following error:

cannot allocate vector of size 399.3 Mb

I have still 12GB RAM free, so I wonder what seems to be the problem...

yeah, that is puzzling ... could you attach/send a .omv file of it?

with thanks

Sure, here you go...

(file is simplified, but the problem already arises with 1 predictor)

hmm, yeah, this is a tricky one. running it here, it does use a surprising amount of RAM (~2 gig).

can you try running the analysis, and watching the memory consumption of the `jamovi-engine` processes (in a task manager/process monitor or some kind)? let us know if it really shoots up, and how high it goes.

this probably going to be a difficult one to resolve, because it will be caused by an R package doing something very inefficiently, and it will involve a) finding out what R package is responsible b) probably a bunch of buck passing between multiple package authors who will ultimately shrug and say "that's just how much memory it uses".

but we never know, we may get lucky.

with thanks

jonathon

should be fixed in the 1.1.6

thanks for reporting

jonathon